Studies of e γx and of e π0x with clas and clas12

Transcription

1 Studies of e γx and of e π0x with clas and clas12 DeepPWG meeting, JLab, June 15, 2017 Hard exclusive photon production BH propagators Possibilities with e γx Extracting PBPT from double spin asymmetry Possibilities with e π 0 X Summary & Conclusions June 15 1

2 Outline 1) e'x -cross section: electron acceptance is relevant for all other measurements cons: we need the acceptance and the luminosity as well as contamination from pions under control. 2) e π 0 X/ e'x ratio (ratio of semi-inclusive pi0 to inclusive electron) For the ratio we just need the gamma acceptance, which could be defined using the KPP cons: need efficiency for photon detection and acceptance under good control 3) e γx/ e'x ratio for the missing mass below \Delta (exclusive photon production) a) In the range of phi (phi<60 degree) it is totally dominated by BH and can be used to extract the ex itself from ratio and control the efficiency of our electron reconstruction. b) It is the same as our DVCS analysis, and can be used as a cross check for extraction with proton detection later on cons: need efficiency for photon detection and acceptance for photons above 3 GeV (easier than for pi0) and well tuned MC to account for pi0 contamination, which is item 2 and can be extracted simultaneously. June 15 2

3 Electroproduction Kinematics γ e θ p γ*->γ require a finite longitudinal momentum transfer defined by the generalized Bjorken variable ξ June 15 3

4 φ-dependent amplitude CBH kinematic factors CInt 5.7 GeV C Int =C BH *x/y*(1+ε 2 ) 2 Strong dependence on kinematics of prefactor φ-dependence, at t t col,p 1 (φ) 0 4 Do the kinematic factors with propagators in T BH and I cancel in the ratio of

5 Azimuthal moments in ep->e p γ Ratio of different contributions to c 0BH BH Interf dvcs c 0L s 1Unp s 1LPol c 1L c0dvcs Different azimuthal moments become relevant in different kinematical regions 5

6 φ-dependent amplitude June 15 6

7 GeV resolution: δp/p~2%, δθ<1 mrad simple box acceptance for photons measure ratios to DIS dvcsgen good agreement with generated BH 7

8 CLAS e1dvcs2 experiment e p γ e Less background than in e1dvcs Large kinematical coverage in x B and t (higher E b ) CLAS@5.9 GeV Calorimeter and superconducting magnet within CLAS torus dedicated calorimeter (424 PbWO4 crystals) detect photons from 5 o -15 o Use e1dvcs2 data to study feasibility of DVCS/BH studies with only electron and photon detection! 8

9 e1dvcs2 distributions Resolution consistent with expectations for CLAS12 9

10 e1dvcs2 distributions Cuts on photon missing mass, minimum energy and the angle between γγ*, may be sufficient, but will require good MC simulation to understand better the background around 180 degrees, where p0 contribution may be significant. DVCS studies with just e and γ are feasible and require only detailed understanding of the photon acceptance (+MC) June 15 10

11 φ-distributions from π 0 s With full exclusivity cut the photons from asymmetric decays may have very different azimuthal distributions compared to symmetric decays, which are used to get the contamination. June 15 11

12 γ MC vs Data Region where BH totally dominates (small t, small photon θ LAB ) Negligible DVCS x-section, small π 0 contamination Rapidly changing prefactors, mainly small φ, hard to detect photons Large angles Uniform coverage in angle φ, photon measurement less challenging DVCS x-section non negligible introduce some model dependence) π 0 dominates the single photon sample (in particular at low Q 2 and large t ) MC Kinematic distributions in x,q 2,t consistent with the CLAS data June 15 12

13 MC studies: BH in A LL eg1dvcs-mc BH-only A LL measured A LL calculated The A LL φ-dependence dominated by BH only at relatively small φ 13

14 Semi-Inclusive electrons (ep->e π 0 X) π 0 The ratio of e π 0 X/e X provides cleanest info on fragmentation function x-bins:0.18 <0.23<0.28< 0.33< 0.38<0.43 Ratios studied using the clasdis LUND generator 1) ratio should have weak dependence on x 2) Ratio should follow z-dependence of the fragmentation function. 14

15 π multiplicities in SIDIS ep e πx Hall-C DSS (Q 2 =2.5GeV 2 ) DSS (Q 2 =25GeV 2 ) π 0 multiplicities less affected by higher twists 0.4<z<0.7 kinematical range, where higher twists are expected to be small 15

16 Suggestions for first publication First measurements could be performed using only identified electrons and photons (e X, e π 0 X, e γx) KPP data can be used to define the efficiency and acceptance for photons in the PCAL+EC Need: Precision calculations of the e X x-sections Precision MC description of the CLAS12 June 15 16

17 summary & plans The eγx with missing mass cuts can be used to study the BH observables (for PbPT,lumi monitoring, ) theoretically under control with high precision dominates all other processes at small φ need dilution factor for gammas (nuclear contributions) Extraction of eγx/ex ratios using first CLAS12 data! minimum effect from charge particle acceptance Extraction of eπ 0 X/eX ratios using first CLAS12 data! provide access to fragmentation functions Plans: Use clas12 software to study eg1dvcs/clas6 e γx/e X Check calculations of the e X x-sections using eγx for BH Compare the A LL with calculations (eg1dvcs data) and develop procedure for extraction of the P B P T from eγx/ex Precision MC description of the CLAS12 comparing with KPP 17

18 support slides June 15 18

19 HT and Semi-Exclusive Pion Production E. Berger, S. Brodsky 1979 (DY), E.Berger 1980, A.Brandenburg, V. Khoze, D. Muller 1995 Fragmentation π + π + π 0 A.Afanasev, C.Carlson, C. Wahlquist Phys.Lett.B398: ,1997 Azimuthal asymmetries with opposite sign from HT effects Effect may be suppressed for semi-exclusive π 0 compared to π +/- Avakian, JLab May 10 19

20 CLAS12 resolutions June 15 20

21 A LL -data vs BH June 15 21

22 φ-dependent ratios (eg1dvcs paper) GPD model independent GPD model dependent (systematics from models) Empty symbols correspond to separate nominator and denominator calculations and filled symbols to calculation of the ratio (right plot is for ratio) Difference ~5% from calculating as June ratio of structure functions

23 φ-dependent amplitude June 15 23

24 φ-dependent amplitude C int /C BH =-x/y*(1+ε 2 ) 2 June The φ-dependence of prefactor doesn t cancel: should be accounted in calculations

25 φ-dependent amplitude June 15 25

26 azimuthal moments in DVCS (example) All moments involve several contributions with different Form Factors and GPDs multiplied by a kinematic term involving propagators June 15 26

27 φ-dependent ratios (eg1dvcs paper) GPD model independent GPD model dependent (systematics from models) June The φ-dependence of prefactor doesn t cancel: should be accounted in calculations

28 June 15 28

29 DVCS: BH propagators BH φ=0 φ=45 φ=90 DVCS x=0.25 Strong dependence on kinematics of prefactor φ-dependence, at t=t col P 1 (φ)=0 require special attention in interpretation of beam averaged beam SSA and in particular x-section differences June Fraction of pure DVCS increases with t and φ

30 Sensitivity on theory input M.Kirch (2006) Dynamical HT may decrease the beam SSA at small t Evolution effects significant at small Q2 June 15 30

31 Contributions to the DVCS cross section VGG BMK BH exact exact Int BH e *DVCS a (BH*DVCS) a VGG VGG x-section described as a sum of moments Azimuthal moments in the DVCS MC (BMK-open symbols) and Vanderhaeghen et al (filled symbols). Reason for difference under studies (2002) June 15 31

32 GPDs from ep->e p γ Requirements for precision (<15%) measurements of GPDs from DVCS SSA: Define the procedure to extract GPDs from A LU effect of finite bins (prefactor variations) ~10% other moments Define background corrections (>1γ) pion contamination ~10% radiative background ADVCS A complete simulation of the whole chain from particle detection to GPD extraction, including the DVCS and background (counts, asymmetries) as well as extraction procedure (averaging over kinematic factors) required to ensure the reliability of measured GPDs. June 15 32

33 Correction factor as a function of bin size φ dependence of in BH and INT terms doesn t cancel for finite bins 0.2<x<0.25 <Int>/<BH> Int(<x,Q>)/BH(<x,Q>) 0.2<x< <x<0.4 φ ( ο ) June 15 33

34 π 0 contamination in epγ-sample CLAS forward Calo single photons from π 0 γ π 0 BH-DVCS γ (photon lab. angle) At large angles detected photon can be used as veto (epx-sample). Cut on the direction of the measured photon (require detection of proton) significantly reduce the contamination (epγ-sample). June 15 34

35 π 0 contamination Main unknown in corrections of photon SSA are the π 0 contamination and its beam SSA. Extract from epπ 0 MC the ratio of single to double photon detection probability MC the ratio for 0 γ MC 1.0<Q 2 <2.0, 0.1<x<0.3 Use epπ 0 MC and data to estimate the contribution of π 0 in the epγ and epx samples Contamination from π 0 photons increasing at large t and x. June 15 35

36 φ-dependence and collinearity cut e16 Collinearity cut y col -y>0.025 eliminates low φ and large t (the beam direction) June 15 36

37 γ MC vs Data Exclusive photon production simulated using a realistic MC(based on S.Korotkov s code) Exclusive photon production simulated using a realistic MC(based on S.Korotkov s code) Kinematic distributions in x,q 2,t consistent with the CLAS data June 15 37

38 GPDs from ep->e p γ Requirements for precision (<10%) measurements of GPDs from DVCS SSA: Define the procedure to extract GPDs from A LU effect of finite bins ~10% Define background corrections pion contamination ~10% radiative background VGG-99 π 0 dominates the single photon sample at low Q 2 in the kinematics where BH is small June 15 38

39 BH cosφ moment BH cosφ moment can generate ~3% sin2φ in the A LU June 15 39